Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Keep perfect predictor for marginsplot

    Dear all,
    I am using mixed and melogit to compute growth curve models. I created a selective sample so there is no variation in outcome at t = 1. In the regular mixed model, this works fine, however, as soon as I switch to melogit (and the depvar is binary) the model drops this first time point and it does no longer show up in marginsplot. I tried the asis option, but then the model does not converge at all. Any ideas what to do here? As an example, attached you find the graph how it is produced from mixed. I would like the same, but from melogit. Of course, at the first point in time, there is no variation.
    Click image for larger version

Name:	bytrack.png
Views:	1
Size:	72.5 KB
ID:	1539797
    Best wishes

    (Stata 16.1 MP)

  • #2
    You didn't get a quick answer. You will increase your chances of a useful answer by following the FAQ on asking questions – provide Stata code in code delimiters, readable Stata output, and sample data using dataex.

    This is unlikely to be a problem someone has addressed before and can give you an easy answer. However, to help you someone would have to cook up data that has your specific problem and estimate it (not knowing if they're estimating it precisely as you have) before they can even begin to solve your issue.

    The inefficient way to solve this problem would be to post the results for margins, convert them into observations and then do the plots. This will allow you to add any observation you want.

    Comment

    Working...
    X